Any component senior executive would simply devolve their military oriented business to the employees that want to execute it.
This failure is acutely felt in Boston Dynamics which was peopled by jingoistic engineers, who were limited in their ability to produce value to Google by the inability of executive to disassociate themselves from the body of that companies work.
Alphabet should have had a second, and majority owned company, take on that drone contract.
Google's "Don't be evil" motto never vanished - it's still in their code of conduct. It got deemphasized from a wordy mention at the top, but its new right-at-the-end placement is still prominent for people who read the whole document (as employees should) or who skip to the end.
Their motto was never "do no evil" - that's a far harder standard given that undesired side effects happen for most actions.
Yup. Though I still don't think it needs a change in intention - i.e. more evil can still result from more power, more "machine learning with more data to assist and suggest to delight users", without any specific evil intent.
They definitely aren't better than most companies in certain corporate matters related to the employee experience (especially for those who are in one or more marginalized demographics), or in their willingness to be a defense contractor at all (back on topic for this news item). They also make plenty of product and PR mistakes.
Still, many tech companies are notably worse on many important issues: more willing to directly facilitate military weaponry and intelligence for lethal purposes, less top-notch at security, fewer and weaker privacy promises, bad at rigorously implementing and adhering to the privacy promises they do make, more in the habit of voluntarily sharing direct access to your personally identifiable data with third parties, less willing to tolerate employees complaining about privacy/security/harassment/workplace issues, etc.
Overall I think Google needs to be watched closely, forced to do the right thing by some mixture of internal external pressure, possibly split into 2-4 smaller companies, and needs to fix certain culture problems .... but I still think it's among the better of the large trendy US tech companies, alongside Microsoft. And GCP is awesome.
Disclosure: I used to work at Google, including the GCP team; my mention of Microsoft is based on second-hand information from people I know who do or did work there.
Overall? You may very well be right. At Google? Based on my memories of the culture there, definitely higher than that. Certainly not 100% though, agreed.
"Don't poison the guests" would be analogous to "don't do [a specific evil thing]." "Don't be evil" is more analogus to "don't get sloppy with hygiene." The second message is a good one for a restaurant to give its workers.
“Don’t be evil” was, from the second it was suggested by (I believe Paul Buchheit) a acknowledgement of failure. The story goes that the, then very young, company was trying to define their values and every suggestion, no matter how broad, was being shot down by a counter-example. Even “be good” was not enough because too vague, or not neutral enough. A young engineer tried to summarise that failure by a double-negative, which left all the moral undefined.
I still think that this problem isn’t resolved: is the problem the existence of violence, its legitimate monopoly by a democratically elected temporary dictator, or that violence is being misused by jingoistic non-sense, and that the US armed forced are controlled by corporate interests? Is evading that question the most moral stance? I feel like it’s mainly an easier one to defend, rather than the better one.
>Google’s... won’t design or deploy AI for "weapons or other technologies whose PRINCIPAL purpose or implementation is to cause or DIRECTLY facilitate injury to people."
This policy is so broad it's meaningless.
PRIMARY purpose - promise us that its primary purpose is something else and we're game.
Direct Injury - AI which determines / designates hostiles and isn't installed directly on the weapon can't cause direct harm.
I'm honestly curious if it was ever about money or just that - if you, you know, ignore the whole "killing people" aspect - killbots are an incredibly difficult and complex and cool piece of technology to work on. You go from problems like "can we draw a box around the faces in this artificial dataset" or "can we keep a car between these white lines" to problems like "can we identity a person from a km away while traveling at 100mph" and "how do you take evasive maneuvers" and "can you do all this with hardware mounted on a drone with an acceptable latency". The whole Jurassic Park "you got so wrapped up in asking if you could you didn't ask if you should" thing.
Which affects shareholders more? The potential negative PR or the potential government contracts?
If it were so cut and dry, I don't think they'd have a policy at all, and I think this policy is the output of a calculated risk assessment where Google determined that the effect of the PR would be greater than the cash.
Plus, there was a chunk of their staff that threatened to leave the company if they went through with their pentagon drone project [0]. I imagine it would be hard to explain that to Wall st.
> ... and I think this policy is the output of a calculated risk assessment where Google determined that the effect of the PR would be greater than the cash.
Huh? It sounds like they're not changing course at all, but just spinning what their doing in a less SkyNet light.
> Which affects shareholders more? The potential negative PR or the potential government contracts?
Considering several of the most valuable companies in the world are U.S. defense contractors, I don't think there's much of question about what affects shareholders more.
My point is that any involvement with war lead to consumer boycotts if Google.
For a company that relies so heavily direct interaction with consumers, involvement with warfare is a huge risk. And the value of the contracts hardly outweigh the risks of backlash against Google primary business.
It's the same reason drug companies work hard to avoid having their drugs used for executions. The value of the contracts is low, compared to their main business which could be threatened by such a move..
Also if you look at trends over the decades it's clear that warfare is increasingly loosing it's glory. People aren't thrilled about murder and mayhem.. sure we'll have military spending for many years to come, but it's likely to loose it's popularity, and eventually fade..
So why is Google even considering to enter this market? When it's only short term and puts Google golden egg at risk? It makes no sense to me.
Translation: They'll do the big bucks basic research and pretend they don't understand the implications of their work or won't bother to consider the unintended consequences.
It's an interesting problem, when the most talented companies in China and Russia will have no problem developing AI weapons for the state to benefit the motherland (and even if they do have a problem with it, they'll have little choice).
My guess is the way the deep state and military industrial complex will deal with this, is to seed new Googles - In-Q-Tel, DARPA, etc - that are more cooperative in this area, and to punish Google in various ways going forward for non-alignment (overtly or otherwise). What use to the military complex is Google if they won't fully cooperate on something like this; time to get themselves a new Google.
Kind of interesting to contrast this with the attitude of earlier generations of technological innovators.
Archimedes developed catapults and canons to fight invading Romans. Leonardo Da Vinci designed war machines, including a giant crossbow. I don’t recall the details, but many European mathematicians in the 18th and 19th centuries rushed to help their countries develop weaponry, which they saw as a sort of patriotic duty. Alan Turing helped invent the computer as part of a military effort. John von Neumann, besides also helping to invent the computer in a military effort, was also an expert in nuclear warheads and work in the Manhattan Project. Richard Feynman worked on ballistics calculations during WWII and then after Pearl Harbir he worked on enriching Uranium for the Manhattan Project. In the 1950s, Gordon Moore (cofounder of Intel) and Robert Noyce created Fairchild Semiconductor, which practically created Silicon Valley and the transistor. The first transistors sold by Fairchild were used in the B70 bomber. Needless to say, the list goes on.
I think that Google is developing weapons, but in the form of spam detectors, face recognition and Geo IP technology. They might not sell them (they don’t need to) but the open-source nature makes using those in an information war far easier.
A lot of Facebook’s effort to track whether Russian agents purchased ads space, how they fostered dissent with controversial groups, etc. all that false under a new military “space” (neither ground, sea or air). Friends of mine who work in security hate the idea of calling it “cyber”, but “information” seems appropriate: cultivating journalists and influencers is key for instance. It’s a war against democratic institutions more than population, but it affect power quite directly.
What Google refused to do was to attach military technology from a previous era (explosive, kinetic weapons) to their own work (face recognition) because they find that abhorrent. That’s actually not the first time: the use of wild animals, poisons, children and later radioactive material and nerve agents in war has been decried by military technologists since Archimedes — generally around the idea of a cleaner violence.
I dont want to say humanity is different, but I think its time to change culture.
I think we went from Hunter Gatherer where land ownership didnt matter to Agriculture where land ownership was everything to enlightenment to post war due to WMDs.
War will kill us all with this new technology, there are people still alive today that were born before Nuclear weapons.
There's something right below the surface that I think the discussions are all glossing over: the difference here is that the United States is conducting wars all over the world with enemies who may threaten, in some conception of the term, American interests, but who present little to no threat to American domestic security. If the US were being invaded I doubt you'd see so much squeamishness.
> "...who may threaten, in some conception of the term, American interests..."
Let's not be naive. "Amwrican interests" is a euphemism for "revenue and potential revenue of America's rich and powerful."
For example, history is clear on Afghanistan. A "conventional war" doesn't work there. But if you want to reap the rewards of all the untapped natural resources __and__ you're not footing the bill for the military's efforts then you simply command your puppet(s) to give it a shot. The only loss is the lives of someone else's kids.
> "Amwrican interests" is a euphemism for "revenue and potential revenue of America's rich and powerful."
I have news for you. In geopolitics, “interests” has always meant primarily economic interests, and economic interests have always been of greatest relevance to the rich. Take the great powers on the eve of World War 1, as an example. Britain had “interests” in Egypt. Russia had “interests” in Turkey. The United States has “interests” in Panama. Millions of people died for these “interests” in WWI. What is meant by “interests” here is business interests. Specifically, Britain seized Egypt in order to control the Suez canal, which was important because it connects the Atlantic and Indian Oceans, and British merchants exported goods to Asia through the canal. All of Russia’s exports had to go through the Turkish straights (controlled by the Ottomans at the time), which connect the Black and Mediterranean Seas. The United States was interested in Panama of course because the Panama Canal was a major conduit of maritime trade between the Atlantic and Pacific Oceans. Such “strategic interests” are the reason behind every war ever fought, it may not be right but it’s nothing unique to the US.
"Every war ever" is overselling it, given other greatest hits like "we should ethnically cleanse this other group of people" and "we should kill the people living on this land and take it for ourselves."
Worth mentioning that da Vinci was a vegetarian animal lover who became obsessed with the destruction of the earth toward the end of his life, repeatedly producing "deluge" drawings of flood waters overtaking the world. He also spent ten months with the Florentine army as preparation for painting the Battle of Anghiari which is pretty clearly an anti-war painting designed to dismay the viewer ("Show someone using his hand as a shield for his terrified eyes... Make the dead, some half buried in the dust, others with the dust all mingled with the oozing blood and changing into crimson mud.")
There's also a theory that he intentionally saboutaged the designs for some of his military machines, although I personally am not convinced by the evidence. [1]
I wish they did more to help the US military with AI technology. It reduces costs of human interpretation and doesn't have the same issues with classified information in the minds of those humans.
There could be great time savings and the potential for higher reliability of image interpretation if unsure AI assessments are passed onto humans for interpretation.
I don't believe that the core of the US military is evil, and as an American I am happy to have it defending my country.
China is a leader in AI and they will develop AI weapons (does anyone doubt this?) Isn't the logical conclusion the country with AI weapons will become militarily superior to one without?
I speak for many people when I say the world doesn't want the U.S. to keep its military superiority. We're tired of their shit. China or U.S. having it, same difference.
Although we are way way off that, a better, more organised, more democratic, more powerful U.N. should have that role.
No. I was referring to the difference between China and the U.S. WW2 is a good point to look at how the two world powers grew to be very different(I intrepid what you were saying was that they are the same) you know the whole 1949 communist thing, cold war into the Korean war, the U.S. having nukes and having just dropped then five years before could do it again but did not. I think that shows something not to negate some of the very horrible things that were done. If we move forward we get to 1989, Tiananmen Square. I'm just saying the two are different and enjoy a mostly open and free internet.
US already lost the benefit of doubt and the excuse of WW2. They literally invented modern terrorism by training the would be terrorist orgonisation leaders against russia during the cold war. Then later invaded afghanistan and iraq for oil, which caused the rise of ISIS. From the world's perspective, US is the biggest culprit in the modern world. Invading other countries and also elected an orange buffoon as their leader.
While the US is far from perfect, it doesn't risk your life or your freedom to speak poorly of the leader of your country as a US citizen. That same cannot be said about a citizen of China. That is the main reason I think it better that the US has the worlds leading military over China.
Ignoring things like democracy, capitalism, socialism, "communism", etc for the moment. China is an ethnostate. Unless you are Han Chinese you will never ever be represented by the Chinese government.
Do you hate the systemic racism, military-industrial complex, environmental abuse, corruption, and overreaching government surveillance present in the US? Yes? Anyone not intentionally being dense can see all those problems are far worse in China. Then consider fact that the Chinese government is actively trying to make those things worse. It fucking infuriates me that anti-American sentiment is so trendy now that people (even as a joke) propose that China might be a better world police.
It doesn't matter. Our indebtedness to China combined with the fact that MNCs lose money in world wars is something of a guarantee that there will never be a war between China and the USA.
I used to have this view; a more precise miliatary has fewer civilian casualties, and shortens wars, limiting overall destruction.
However, my mind was changed by the idea that if killing people is easy, and there is little risk to the operators of the equipment, it will be done all the more. Countries will go to war much less overall, if there are higher stakes. I like there being high stakes.
> However, my mind was changed by the idea that if killing people is easy, and there is little risk to the operators of the equipment, it will be done all the more. Countries will go to war much less overall, if there are higher stakes. I like there being high stakes.
Modern warfare is not conducted like this. Killing is already easy but that’s not how wars are won. The trend of warfare since WWII is on precision targeting. Take a look at munitions technology and systems from then to today. The US military wants to minimize casualties and collateral damage while acheiving mission objectives.
Google should be using AI technology to save the lives of American soldiers.
It was Albert Einstein who signed a letter (ghost written by physicist Leo Szilard) written to Roosevelt about suggesting initiating what became the Manhattan project which created the nuclear weapons that saved an estimated 500,000 American soldiers lives when used on Japan.
Was it wrong of President Truman to save the lives of 500,000 American soldiers?
I'm sure all those people who at Google who forced this decision will feel great. When the tech industry in China helps China become a pre-immenient force using AI for military purposes, I'm sure they will regret their decision.
Is there a competition between China and the US to drone the Middle East I don’t know about? I certainly won’t feel badly when the technology is instead developed by another company and used to bomb non-combatants in Pakistan.
It’s not miliatary contracts I have a problem with, it’s this particular program, this particular method.
>>> AI applications for weapons and technologies that "gather or use information for surveillance,"
Couldn't be more ironic. Google's core business is surveillance and serving ads - all aided by ai.
This whole thing is a joke - a bunch of know-it-all overrated and overpayed jerks pretending to know what's best for the world/country. Why and how is it assumed that they know better than lawmakers, government officials and security agencies?
At the very least, this is a vote of no-confidence against the very institutions and people instilled to defend the country and further it's interests!!
When Motorola was acquired by Google, plenty of googlers commented on how much more smarter the average Googler was. And how the new genius Google upper brass world turn Motorola into a money making machine. However, they failed miserably at building and selling good hardware, and that struggle continues to this day through their pixel, Nexus and HTC devices. So - a bunch of cocky software engineers couldn't even build and sell better phones, why'd they be right on national security?
54 comments
[ 4.0 ms ] story [ 117 ms ] threadThis failure is acutely felt in Boston Dynamics which was peopled by jingoistic engineers, who were limited in their ability to produce value to Google by the inability of executive to disassociate themselves from the body of that companies work.
Alphabet should have had a second, and majority owned company, take on that drone contract.
Their motto was never "do no evil" - that's a far harder standard given that undesired side effects happen for most actions.
One's capacity for evil indeed grows greatly when one is given too much power.
Likewise, as Google's power has grown inordinately, their public emphasis on "doing no evil" has diminished
They definitely aren't better than most companies in certain corporate matters related to the employee experience (especially for those who are in one or more marginalized demographics), or in their willingness to be a defense contractor at all (back on topic for this news item). They also make plenty of product and PR mistakes.
Still, many tech companies are notably worse on many important issues: more willing to directly facilitate military weaponry and intelligence for lethal purposes, less top-notch at security, fewer and weaker privacy promises, bad at rigorously implementing and adhering to the privacy promises they do make, more in the habit of voluntarily sharing direct access to your personally identifiable data with third parties, less willing to tolerate employees complaining about privacy/security/harassment/workplace issues, etc.
Overall I think Google needs to be watched closely, forced to do the right thing by some mixture of internal external pressure, possibly split into 2-4 smaller companies, and needs to fix certain culture problems .... but I still think it's among the better of the large trendy US tech companies, alongside Microsoft. And GCP is awesome.
Disclosure: I used to work at Google, including the GCP team; my mention of Microsoft is based on second-hand information from people I know who do or did work there.
I really wonder what is the percentage of employees that read their company's code of conducts. I wouldn't be surprised if it was 5 %.
Imagine that Disney used that motto, would you still send your children to DisneyLand?
Or a restaurant saying "don't poison the guests", would you still eat there?
And yes, I'd like a restaurant to encourage its workers to take steps to ensure the food is safe.
By telling them to not poison the guests?
Do I really have to explain what's wrong here?
I still think that this problem isn’t resolved: is the problem the existence of violence, its legitimate monopoly by a democratically elected temporary dictator, or that violence is being misused by jingoistic non-sense, and that the US armed forced are controlled by corporate interests? Is evading that question the most moral stance? I feel like it’s mainly an easier one to defend, rather than the better one.
This policy is so broad it's meaningless.
PRIMARY purpose - promise us that its primary purpose is something else and we're game.
Direct Injury - AI which determines / designates hostiles and isn't installed directly on the weapon can't cause direct harm.
It sounds like the want to keep selling cloud services, which is pretty much a commodity at this point.
Even if weak, any policy is better than none.
Anyways, given Googles consumer brand I suspect any form of involvement with weapons systems would be unwise :) I'm surprised they even considered it.
Reason and morality falls apart when confronted with shareholder value.
If it were so cut and dry, I don't think they'd have a policy at all, and I think this policy is the output of a calculated risk assessment where Google determined that the effect of the PR would be greater than the cash.
Plus, there was a chunk of their staff that threatened to leave the company if they went through with their pentagon drone project [0]. I imagine it would be hard to explain that to Wall st.
0 - https://www.bloomberg.com/news/articles/2018-06-01/google-wo...
Huh? It sounds like they're not changing course at all, but just spinning what their doing in a less SkyNet light.
> Which affects shareholders more? The potential negative PR or the potential government contracts?
Considering several of the most valuable companies in the world are U.S. defense contractors, I don't think there's much of question about what affects shareholders more.
Better to do a spin-off company instead..
For a company that relies so heavily direct interaction with consumers, involvement with warfare is a huge risk. And the value of the contracts hardly outweigh the risks of backlash against Google primary business.
It's the same reason drug companies work hard to avoid having their drugs used for executions. The value of the contracts is low, compared to their main business which could be threatened by such a move..
Also if you look at trends over the decades it's clear that warfare is increasingly loosing it's glory. People aren't thrilled about murder and mayhem.. sure we'll have military spending for many years to come, but it's likely to loose it's popularity, and eventually fade..
So why is Google even considering to enter this market? When it's only short term and puts Google golden egg at risk? It makes no sense to me.
In other words: the status quo.
My guess is the way the deep state and military industrial complex will deal with this, is to seed new Googles - In-Q-Tel, DARPA, etc - that are more cooperative in this area, and to punish Google in various ways going forward for non-alignment (overtly or otherwise). What use to the military complex is Google if they won't fully cooperate on something like this; time to get themselves a new Google.
Archimedes developed catapults and canons to fight invading Romans. Leonardo Da Vinci designed war machines, including a giant crossbow. I don’t recall the details, but many European mathematicians in the 18th and 19th centuries rushed to help their countries develop weaponry, which they saw as a sort of patriotic duty. Alan Turing helped invent the computer as part of a military effort. John von Neumann, besides also helping to invent the computer in a military effort, was also an expert in nuclear warheads and work in the Manhattan Project. Richard Feynman worked on ballistics calculations during WWII and then after Pearl Harbir he worked on enriching Uranium for the Manhattan Project. In the 1950s, Gordon Moore (cofounder of Intel) and Robert Noyce created Fairchild Semiconductor, which practically created Silicon Valley and the transistor. The first transistors sold by Fairchild were used in the B70 bomber. Needless to say, the list goes on.
A lot of Facebook’s effort to track whether Russian agents purchased ads space, how they fostered dissent with controversial groups, etc. all that false under a new military “space” (neither ground, sea or air). Friends of mine who work in security hate the idea of calling it “cyber”, but “information” seems appropriate: cultivating journalists and influencers is key for instance. It’s a war against democratic institutions more than population, but it affect power quite directly.
What Google refused to do was to attach military technology from a previous era (explosive, kinetic weapons) to their own work (face recognition) because they find that abhorrent. That’s actually not the first time: the use of wild animals, poisons, children and later radioactive material and nerve agents in war has been decried by military technologists since Archimedes — generally around the idea of a cleaner violence.
I think we went from Hunter Gatherer where land ownership didnt matter to Agriculture where land ownership was everything to enlightenment to post war due to WMDs.
War will kill us all with this new technology, there are people still alive today that were born before Nuclear weapons.
Let's not be naive. "Amwrican interests" is a euphemism for "revenue and potential revenue of America's rich and powerful."
For example, history is clear on Afghanistan. A "conventional war" doesn't work there. But if you want to reap the rewards of all the untapped natural resources __and__ you're not footing the bill for the military's efforts then you simply command your puppet(s) to give it a shot. The only loss is the lives of someone else's kids.
I have news for you. In geopolitics, “interests” has always meant primarily economic interests, and economic interests have always been of greatest relevance to the rich. Take the great powers on the eve of World War 1, as an example. Britain had “interests” in Egypt. Russia had “interests” in Turkey. The United States has “interests” in Panama. Millions of people died for these “interests” in WWI. What is meant by “interests” here is business interests. Specifically, Britain seized Egypt in order to control the Suez canal, which was important because it connects the Atlantic and Indian Oceans, and British merchants exported goods to Asia through the canal. All of Russia’s exports had to go through the Turkish straights (controlled by the Ottomans at the time), which connect the Black and Mediterranean Seas. The United States was interested in Panama of course because the Panama Canal was a major conduit of maritime trade between the Atlantic and Pacific Oceans. Such “strategic interests” are the reason behind every war ever fought, it may not be right but it’s nothing unique to the US.
Thanks. I'm aware. It's the majority of the proles who are fooled by the masking of the euphemism, and the massive hypocrisy it creates.
The point being, it's __not__ about liberty, freedom, human rights, etc. It __is__ about economic and cultural imperialism.
The phrase is exactly the kind of language abuse that Orwell so kindly warned us about.
There's also a theory that he intentionally saboutaged the designs for some of his military machines, although I personally am not convinced by the evidence. [1]
[1] https://www.telegraph.co.uk/news/uknews/1415660/Leonardo-da-...
I wish they did more to help the US military with AI technology. It reduces costs of human interpretation and doesn't have the same issues with classified information in the minds of those humans.
There could be great time savings and the potential for higher reliability of image interpretation if unsure AI assessments are passed onto humans for interpretation.
I don't believe that the core of the US military is evil, and as an American I am happy to have it defending my country.
Although we are way way off that, a better, more organised, more democratic, more powerful U.N. should have that role.
I think there's a big difference here.
However, my mind was changed by the idea that if killing people is easy, and there is little risk to the operators of the equipment, it will be done all the more. Countries will go to war much less overall, if there are higher stakes. I like there being high stakes.
Modern warfare is not conducted like this. Killing is already easy but that’s not how wars are won. The trend of warfare since WWII is on precision targeting. Take a look at munitions technology and systems from then to today. The US military wants to minimize casualties and collateral damage while acheiving mission objectives.
It was Albert Einstein who signed a letter (ghost written by physicist Leo Szilard) written to Roosevelt about suggesting initiating what became the Manhattan project which created the nuclear weapons that saved an estimated 500,000 American soldiers lives when used on Japan.
Was it wrong of President Truman to save the lives of 500,000 American soldiers?
It’s not miliatary contracts I have a problem with, it’s this particular program, this particular method.
Couldn't be more ironic. Google's core business is surveillance and serving ads - all aided by ai.
This whole thing is a joke - a bunch of know-it-all overrated and overpayed jerks pretending to know what's best for the world/country. Why and how is it assumed that they know better than lawmakers, government officials and security agencies?
At the very least, this is a vote of no-confidence against the very institutions and people instilled to defend the country and further it's interests!!
When Motorola was acquired by Google, plenty of googlers commented on how much more smarter the average Googler was. And how the new genius Google upper brass world turn Motorola into a money making machine. However, they failed miserably at building and selling good hardware, and that struggle continues to this day through their pixel, Nexus and HTC devices. So - a bunch of cocky software engineers couldn't even build and sell better phones, why'd they be right on national security?
Call it the Joshua Project. :)